Recently added
Plugins grouped by the day they were added to the directory, newest first.
Sep 16, 2026
- shipwright-compliance13stars
Generate audit-ready compliance documentation from Shipwright pipeline data (RTM, test evidence, change history, SBOM)
Claude Code - shipwright-changelog13stars
Generate changelogs from Conventional Commits, create tags, and manage PRs
Claude Code - shipwright-build13stars
Implement code from /shipwright-plan sections with TDD, code review, Conventional Commits, and git workflow
Claude Code - shipwright-adopt13stars
Onboard an existing repository (brownfield) into the Shipwright SDLC — analyze codebase, generate docs, configs, compliance artifacts, and an E2E baseline
Claude Code - draft-detective21stars
Document review skills from Draft Detective: AI-powered checks that help authors catch issues in their documents before publication.
CodexClaude Code27 skills - slipway0stars
slipway implements the Agentic Delivery Lifecycle (ADLC): interview-driven repo intake, app classification, hardened Dockerfiles, Terraform on Azure, GitHub Actions CI/CD with immutable tags, deployment verification and Jira tracking. Every platform choice is an interview option.
Claude Code8 skills3 MCP servers - craft0stars
An evolvable Agent runtime with context, capability discovery, verified execution, quality evidence, and bounded workflow evolution.
Codex1 skill1 MCP server - craft-workflow-evolution0stars
Compatibility helper that turns evidence-backed execution observations into bounded Workflow drafts.
Codex1 skill1 MCP server - craft-skill-quality0stars
Compatibility alias for Craft Quality; evaluates versioned Agent assets with repeatable evidence.
Codex1 skill1 MCP server - craft-quality0stars
Standalone repeatable evaluation and promotion evidence for Agent assets and task outcomes.
Codex1 skill1 MCP server - craft-memory0stars
Standalone scoped and evidence-linked memory for AI hosts.
Codex1 skill1 MCP server - craft-knowledge0stars
Standalone governed knowledge for people and AI hosts.
Codex1 skill1 MCP server - craft-context0stars
Standalone governed knowledge, scoped memory, and reproducible context for AI hosts.
Codex1 skill1 MCP server - craft-capability0stars
Standalone discovery and governance for Skills, MCP tools, Workflows, and adapters.
Codex1 skill1 MCP server - codex-celebrate5stars
Native full-screen confetti when the main Codex agent stops responding.
CodexClaude Code - team-standards1stars
团队 AI 编程标配：编码规范（含安全红线）与架构规范（Go）、Go 开发规范（godev）、后端/前端测试规范、规格 skills（techspec、sddspec、tech-diagram-design）、团队知识库（kb skill + kb-reviewer + /kb-health）、提测命令、SDD 交付循环（/devloop + test-writer/reviewer/implementer 三角色，contract 按行为编号逐条收敛）
Claude Code9 skills - inbound-sunum0stars
Inbound ajans standardında SEO/GEO performans değerlendirme sunumu (PPTX) üretir - aylık, çeyreklik, yarıyıl ve etki/migrasyon analizi modlarında. Inbound Design System görselini, icerik-dili-rehberi dil standardını ve GSC/GA4/Keyword Planner/SEOmonitor/Ahrefs veri şemalarını uygular. Üretime başlamadan önce sunumun amacını, dönemini ve hangi verilerin dahil edileceğini sorar; elde olmayan export'ları nasıl alınacağıyla birlikte net olarak ister.
Claude Code1 skill - verification-hooks1stars
DEPRECATED (2026-09-16) — hooks removed; plugin is a no-op. Independent verification lives in dev-process. See README.
Claude Code1 skill - total-recall1stars
Front door for Total Recall — Kyle's externalized self-model. Provides skill modes (status, search, ingest, brief) and the /tr-brief cold-start command. Backend at ~/projects/total-recall/.
Claude Code1 skill - tmux1stars
Tmux session control for Claude Code — launch test sessions, send commands, read output with multi-agent safety
Claude Code1 skill - session-historian1stars
Read and analyze Claude Code session history for debugging, continuity, and workflow analysis
Claude Code1 skill - secret-vault1stars
Intercept and securely store secrets pasted in chat. Auto-redacts from session logs.
Claude Code1 skill - pre-mortem1stars
Pre-mortem analysis skill — define success/failure criteria before execution. Forces agents to articulate what bad output looks like before starting multi-step work. Based on Gary Klein's prospective hindsight research.
Claude Code1 skill - handoff1stars
Save session context and resume in fresh Claude sessions without losing continuity
Claude Code1 skill - gastown1stars
Multi-agent orchestration system for coordinating Claude Code instances across projects. Includes skills for Gas Town environment, gt/bd commands, convoys, molecules, and the propulsion principle.
Claude Code2 skills - friction1stars
Friction mining lifecycle for Total Recall — mining session transcripts, writing v2 stories, extracting moments from Kyle's comments, logging to the knowledge graph, and enriching over time. Encodes quality conventions so future sessions don't need re-teaching.
Claude Code1 skill - fable-mode1stars
Fable 5's judgment habits as skills + hooks for Opus 4.8 — precondition checking, executive-mode interaction, one-screen output, adversarial self-review, and two enforcement hooks (rendered-proof, session-exit-hygiene)
Claude Code4 skills - educate-me1stars
Empathetic 1-on-1 tutor that teaches any subject using a structured MECE framework
Claude Code1 skill - dev-process1stars
Structured development pipeline with spec-driven architecture, review gates, independent verification, and agent team coordination. Right-sizes the process (Tier S/M/L) and enforces an extensible Verification Suite of gates via the verify-before-close hook.
Claude Code - deploy-standard1stars
Enforces Kyle Deploy Standard v1 — all deployable projects use /opt symlink layout, Docker, /mnt/pi-data bind mounts.
Claude Code1 skill - capture1stars
Quick-capture tasks, bookmarks, ideas, follow-ups, and research items to a personal inbox via the capture-mcp MCP server.
Claude Code1 skill - capa1stars
Corrective and Preventive Action (CAPA) process skill. Investigates process failures, researches best practices, fixes the process first, then re-executes to fix the original defect. Includes SQLite tracking database for pattern analysis across incidents.
Claude Code1 skill - delivery0stars
Implement, review, and deliver an approved plan slice by slice through Worker and Reviewer subagents: a test gate on the Worker, a read-only Reviewer, and a human-approval prompt for commands that deploy, publish, change shared state, or destroy work.
Claude Code1 skill - architecture-lead0stars
Principal-level architecture consultant subagent team that runs a gated discovery-to-brief workflow — with a skeptic/practical-planner check-and-balance and a cartographer that diagrams every architectural output — without touching implementation.
Claude Code3 skills - tool-timing-badge32kstars
Times every tool call and wraps the engine's ToolUse rendering with a colored duration badge from the surface's own element table, on the terminal, Desktop and mobile. Hooks tool.call and ui.render using JSX.
Claude Code - secret-redactor32kstars
Redacts API keys, tokens, JWTs, private keys and connection strings from every tool result before the model reads them, notes the redaction through context, and refuses to echo a redacted placeholder back into a Bash command. A tool.call hook in the after placement.
Claude Code - protected-paths-guard32kstars
Denies Edit, Write and NotebookEdit calls on sensitive paths (.env files, lockfiles, CI workflows, git internals, private keys) with a glob allowlist override. A tool.call hook with an array matcher.
Claude Code - large-edit-confirmation32kstars
Asks the user, in the engine's AskUserQuestion dialog ($.ui.ask), before Claude edits or overwrites a file larger than a configurable line count; denies by default when nobody can answer (headless runs). A tool.call hook on Edit and Write using $.fs.read.
Claude Code - block-destructive-commands32kstars
Denies Bash commands that match destructive patterns (recursive rm on roots, force push, hard reset, destructive SQL, disk formatting) before they run. A tool.call hook with a Bash matcher that returns { deny } without calling next.
Claude Code - webfetch-cache32kstars
Short-circuits repeated WebFetch calls for the same URL and prompt within a session, answering { result } from cache without a network round trip. A tool.call hook with TTL and size limits.
Claude Code - npm-to-pnpm-rewriter32kstars
Rewrites npm and npx invocations to pnpm (or yarn / bun) before the Bash tool runs them and tells the model what ran through context. A tool.call hook that forwards a modified event (the modifying placement).
Claude Code - universal-audit-log32kstars
One hook on * that records every event on $ (tool calls, prompts, turns, other plugins' fs/http/process calls) as JSON lines with origin plugin and tier, duration and outcome, including denials. Buffered and flushed to a JSONL file through $.fs at the end of every turn.
Claude Code - websearch-to-exa32kstars
Overrides the built-in WebSearch tool and routes queries to the Exa search API through $.http.fetch, answering in WebSearch's own result shape, with automatic fallback to the built-in search. A tool.call hook that replaces a core tool.
Claude Code - typing-test32kstars
Typing test above the Claude Code prompt: retype a line of code as fast as you can. Plays while Claude works, pauses when the turn ends, keeps your best score; zero tokens. A single game split out of cc-arcade (MIT, Seza Akgün). Needs function hooks (early access) and an interactive terminal with a mouse.
Claude Code - tool-defense32kstars
A tower defense above the Claude Code prompt where every enemy is one of Claude's real tool calls: Bash sends runners, an Edit or Write a slow tank, a web or MCP call a flyer, an Agent a boss. Build and upgrade towers by clicking beside the road, hold the base through Claude's turn. Zero tokens. Needs function hooks (early access) and an interactive terminal with a mouse.
Claude Code - tetris32kstars
Tetris above the Claude Code prompt: fit the falling pieces and clear lines. Plays while Claude works, pauses when the turn ends, keeps your best score; zero tokens. A single game split out of cc-arcade (MIT, Seza Akgün). Needs function hooks (early access) and an interactive terminal with a mouse.
Claude Code - snake32kstars
Snake above the Claude Code prompt: eat, grow, and keep off the walls and yourself. Plays while Claude works, pauses when the turn ends, keeps your best score; zero tokens. A single game split out of cc-arcade (MIT, Seza Akgün). Needs function hooks (early access) and an interactive terminal with a mouse.
Claude Code - pong32kstars
Pong above the Claude Code prompt: first to 5 against the computer. Plays while Claude works, pauses when the turn ends, keeps your best score; zero tokens. A single game split out of cc-arcade (MIT, Seza Akgün). Needs function hooks (early access) and an interactive terminal with a mouse.
Claude Code - pet32kstars
A virtual pet above the Claude Code prompt that grows as Claude works: passing tests, commits and edits grant XP and mood, failing tests cost mood, idling drains it; egg → baby → kid → adult → legend. Zero tokens. Split out of cc-arcade (MIT, Seza Akgün). Needs function hooks (early access) and an interactive terminal.
Claude Code - pacman32kstars
Pac-Man above the Claude Code prompt: a 19×15 maze, four ghosts that chase, scatter and flee, power pellets, three lives and faster levels. Plays while Claude works, pauses when the turn ends, keeps your best score; zero tokens. Needs function hooks (early access) and an interactive terminal.
Claude Code